hlmhmm metacognition seed selection and model comparison

library(reticulate)
use_condaenv("glmhmm")

source_python("python/glm_hmm.py")

import numpy as np
import numpy.random as npr
import matplotlib.pyplot as plt
import ssm 

quit

library(tidyverse)
library(broom)
library(dplyr)
library(R.matlab)
library(cowplot)

list.files("R", full.names = TRUE) |> 
  walk(source)

load("RData/glm_hmm1_seeds.RData")
load("RData/glm_hmm2_seeds.RData")
load("RData/glm_hmm3_seeds.RData")

Renames the states according to their values (so similar states between subjects can be identified)

# glm_hmm1 <- glm_hmm1_seeds %>% 
#     mutate(fit = list(glm_hmm_extract_best(fit))) %>%
#     order_states_glm_hmm()
# 
# glm_hmm1
# 
# save(glm_hmm1, file = "RData/glm_hmm1.RData")

# glm_hmm2 <- glm_hmm2_seeds %>% 
#   mutate(fit = list(glm_hmm_extract_best(fit))) %>%
#   order_states_glm_hmm()
# 
# glm_hmm2
# 
# save(glm_hmm2, file = "RData/glm_hmm2.RData")

# glm_hmm3 <- glm_hmm3_seeds %>%
#   mutate(fit = list(glm_hmm_extract_best(fit))) %>%
#   order_states_glm_hmm()
# 
# glm_hmm3
# 
# save(glm_hmm3, file = "RData/glm_hmm3.RData")

load("RData/glm_hmm1.RData")
load("RData/glm_hmm2.RData")
load("RData/glm_hmm3.RData")

———————————————————————

MODEL COMPARISON

---

liks de cada modelo

DL:

liks_glm_hmm <- liks_glm_hmm1 |> 
  bind_rows(liks_glm_hmm2) |> 
  bind_rows(liks_glm_hmm3) |> 
  mutate(aic = -2 * log_lik + 2 *n_par)

save(liks_glm_hmm, file = "RData/liks_glm_hmm.RData")

liks_glm_hmm_best <- liks_glm_hmm |> 
  group_by(subject) |> 
  filter(aic == min(aic)) 

liks_glm_hmm_best |> 
  ungroup() |> 
  count(model)

save(liks_glm_hmm_best, file = "RData/liks_glm_hmm_best.RData")